Suppose you obtain data on a sample of working adult individuals in the U.S. (indexed by i = 1, ..., N), who self-reported the following variables:

labinci = annual wage and salary income from labor (after taxes)

nonlabinci = annual non-labor income (after taxes)

hoursi = annual hours worked

educi = total years of completed schooling

agei = age in years

malei = a dummy variable equal to 1 for males, 0 for females

2.1 Using these data, how would you estimate the wage elasticity of labor supply? Write down a specific regression, and feel free to define any new variables using the ones you already have, if you think they would be relevant. Make sure to mention what your coefficient of interest is.

2.2 How would you test each of the following three hypotheses: A) the substitution effect dominates the income effect. B) leisure is a normal good. 2 C) all else equal, men work significantly more hours than women. Write down specific null (H0) and alternative (H1) hypotheses using parameters from the regression equation you wrote down in 2.1.

2.3 Suppose you estimate an elasticity of 0.2. How would you interpret this?

2.4 Suppose you are particularly worried about people misreporting their annual work hours. Do you think the true labor supply elasticity would be larger or smaller than 0.2? Explain.

2.5 Besides measurement error, what other potential concerns do you have about this regression approach (e.g. with respect to the key OLS assumptions being potentially violated)? Which slope coefficients do you think may be biased? Explain.